Yes, reliable data ingestion often involves Kafka, which can feel complex. An alternative is the transactional COPY INTO approach used by platforms like Snowflake and Databend. This command supports "exactly-once" ingestion, ensuring data is fully loaded or not at all, without requiring message queues or extra infrastructure.

https://docs.databend.com/sql/sql-commands/dml/dml-copy-into...